Does the Teacher Need Another Assistant? Education-Focused AI "Sanna" Launches in Europe

Does the Teacher Need Another Assistant? Education-Focused AI "Sanna" Launches in Europe

AI Can Increase Educational Materials, But Can It Increase Teachers' Time? "Sanna" Questions the Future of Schools

After classes end, preparation for the next day's lessons begins. Teachers rearrange explanations for students who need more time to understand and prepare additional assignments for those who can move ahead. The more teachers try to tailor to each child, the more one type of material becomes insufficient.

As a tool to support this work, a new AI for teachers has emerged from Europe. It's important to focus not only on "what can be created" but also on how this tool will change the daily life of teachers.


What is "Sanna," the AI Supporting Teachers' Work?

On October 1, 2026, Sanoma Learning announced the AI assistant "Sanna." It supports lesson planning, creating additional materials, adjusting materials according to learning stages, and providing feedback.

According to the company, it was developed based on curricula from various countries and the company's educational content, with trials involving over 1,500 teachers across Europe. They are offering a 30-day trial in the Netherlands, Spain, Italy, Poland, Belgium, Finland, and Sweden. Japan is not included in the target countries for this announcement.

The output is to be verified and edited by teachers before use. These features and conditions are based on the company's announcement and should be read separately from independently verified educational effectiveness research results.


Between "Creating Materials" and "Usable in Class"

Even if AI can create problems quickly, it doesn't necessarily mean they are suitable for the children in front of you.

For example, what if a problem assumes concepts not yet taught in Japanese elementary school math? Even if the answer is correct, it might be difficult to use in that day's lesson. If the language in a word problem is too difficult, it might measure reading comprehension differences rather than calculation skills.

What teachers need is not only accuracy in content but also materials that include teaching order, explanation granularity, and consideration for common stumbling points.

From this perspective, the value of educational AI is not determined solely by its ability to produce large amounts of text. Its practicality depends on how few corrections are needed to fit the conditions of "using this explanation for this grade and this unit."

However, assurance cannot be based solely on claims of curriculum compatibility. Teachers must verify the actual output and reflect on its use in the classroom.


High Ratings from Trial Users Are Not Proof of Improved Academic Performance

Sanna's official site reports that in a trial survey, 89% of users rated the quality of the output as equal to or better than self-made materials. Additionally, 81% of teachers used the materials created during the trial in their classes.

These figures indicate the potential acceptance of the tool for creating educational materials. However, these are evaluations from trial users disclosed by the providing company and are not numbers that demonstrate improved academic performance of students.

The feeling of convenience by teachers, the use of materials in classes, and the deepening of children's understanding are related but not the same. When considering the introduction effect, it's necessary to measure each separately.

For example, even if the creation time is shortened, if the same amount of time is spent on content revisions, the reduction in burden is small. On the other hand, even if the required time does not change, the ability to create supplementary materials that were previously unavailable could provide educational value.

It is important to consider both "how much time was saved" and "what support became possible."


Voices of Expectation on Social Media, But Not Post-Release Reviews

On public social media, reactions to Sanoma Learning's pre-release LinkedIn post about the AI for teachers showed expectations for AI rooted in education and centered around teachers.

For example, Jorge Aguirre Cabrera welcomed the approach of advancing educationally responsible, teacher-centered AI. Marco Residori expressed congratulations on the release and anticipation to try it. These are summaries of the post contents.

However, these are comments on pre-release announcements. They are not post-release evaluations reporting "overtime has decreased" or "classes have improved" from continued use of Sanna. Also, the few reactions gathered on the company's official post cannot be treated as the opinion of all teachers.

What can be read at this stage is the expectation for an assistant designed for education, not the final evaluation of its outcomes. Moving forward, specific reports on which tasks helped teachers and what corrections were necessary will become important.


European Teachers Also Distinguish Between AI Convenience and Educational Effectiveness

In a survey announced by Sanoma Learning in September 2026, targeting over 20,000 teachers from 14 European countries, 16% of teachers believed that general-purpose AI improves learning outcomes. Meanwhile, the percentage of those who answered that AI used in education should be designed for educational purposes reached 75-93% depending on the country.

This survey was also conducted by the company, with data collection handled by GfK. While it's best to avoid directly applying these numbers to all European teachers, it suggests a distance between expectations for tools and confidence in educational effectiveness.

In Japan as well, separating evaluations as a convenient tool for creating educational materials and as a means to improve children's learning can be a starting point to avoid excessive expectations.


In Japan, the Question Is Whether "Work Will Decrease After Introduction"

For Japanese schools, teachers' time is a pressing issue.

According to the results of TALIS 2024 presented by the Ministry of Education, Culture, Sports, Science and Technology, the weekly working hours of Japanese teachers are 52.1 hours for elementary schools and 55.1 hours for junior high schools. Although decreased from the previous survey, both are considered the longest among participating countries. While it is a questionnaire survey and attention to differences in systems and cultures is necessary, it serves as material for considering the magnitude of work burden.

If AI is to be introduced in such environments, it is necessary to look at the entire work after introduction, not just the generation function.

Logging in procedures increase. Materials are transferred to another system. Training on how to use it is received. Outputs are verified. Usage results are reported. Even individually small tasks can accumulate to become a new burden.

In trials, it is insufficient to measure only the time taken to create materials. It is necessary to compare the total time including verification, revision, printing, and sharing, and also investigate differences by subject and grade. Listening to the reasons from teachers who stopped using it, not just those who found it convenient, can lead to improvements.


Three Conditions to Consider for Application in Japan

First, it must be compatible with Japanese educational content.

Even if a service for Europe is compatible with the curricula of various countries, it does not automatically align with Japan's educational guidelines or adopted textbooks. Being able to respond in Japanese and being usable in Japanese classes are also different. If considering a similar system domestically, design including the conditions for using materials is necessary.

Second, clarify the responsibility for inputting information and verification.

The Ministry of Education, Culture, Sports, Science and Technology's Generative AI Guidelines Ver.2.0 emphasize human-centered utilization, with humans making the final judgment and taking responsibility. It also shows consideration for personal information, privacy, and copyright.

In situations where teachers at each school are left to make decisions on their own, it is difficult to use with confidence. The introducer needs to specify how much student information can be input, who verifies the generated content, and who to consult when in trouble.

Third, the use of the time saved.

If faster material creation results in "let's increase the number of handouts," teachers will not find it easier. Schools should also reconsider their operations so that the time freed by AI can be directed towards rest or interaction with children.


Can We Increase Teachers' Free Time?

The emergence of Sanna prompts consideration of where AI should be placed in teachers' work.

For example, while AI can support creating alternative explanations, the decision of what words to say to a child today involves daily relationships and observations. The ability to prepare materials and the ability to notice changes in children are both necessary.

When evaluating AI for teachers in Japan, it is important not to measure success solely by the number of printed sheets generated. Did teachers get home earlier? Did they have time to reflect on their lessons? Were they able to approach a child who hesitated to ask questions?

The value of introduction needs to be confirmed not only by the output on the screen but also by the time created in the classroom.


Source URL

Related Articles